ansys 14.0 fluid dynamics update - dipankar choudhury

© 2011 ANSYS, Inc. October 24, 20111

ANSYS Fluid Dynamics Update

Dipankar ChoudhuryVP, Research


ANSYS Fluid Dynamics

ANSYS Fluid Dynamics Products• ANSYS FLUENT• ANSYS CFX• ANSYS CFD‐Post• ANSYS TurboGrid• ANSYS Polyflow

ANSYS Fluids Dynamics at 14.0• Significant benefits for users

• Including geometry and meshing• Look for detailed materials at release time


Product Creation Teams (~600 Members)

CentralDevelopment Unit

Brian Drew•Framework, Tools•Geometry/CAD•Meshing•Engineering KnowledgeManager•Design Explorer•Dev Services•Funded development~30% dev. staff

Technologist:Chris Hawkins

Product Mgmt: Scott Gilmore

FluidsSystems

Nelson Carter•FLUENT, CFX, common platform•Specialty CFD•R&D initiatives•Funded development~25% dev. staff

MechanicalSystems

Joe Manich•Mechanical•Explicit•Multibody dyn.•R&D initiatives•Funded development~25% dev. staff

Physics Business UnitJoe Solecki

Technologists: Dave Conover, Michael Engelman, Dipankar ChoudhuryProduct Management: André Bakker

Electronics Business Unit

Shane Emswiler•High Frequency•Low Frequency•Electro-mechanical•Signal Integrity•RF/Electronics•Funded development•R&D initiatives~ 20% dev. staff

Dev. Mgmt:Nancy LambertProduct Mgmt: Larry Williams

Product Strategy & Planning, Partnerships, Corporate Product ManagementJosh Fredberg, S. Subbiah, and Todd McDevitt


We remain committed to developing and releasing an integrated, unified fluids product• That will draw upon the best of both FLUENT and CFX, as well as add new functionality

• That will minimize migration hurdles for our fluids customers

The first release of this product is targeted for ANSYS R15.0

FLUENT and CFX will continue to be developed in the meantime (and actively maintained for a few releases beyond R15.0)• To ensure smooth migration for our fluids customers

Our Fluids Strategy


• Rapid & Robust Meshing

• Workflow & Usability

• Multiphysics and Systems Coupling

• Solver and HPC Performance

• Rotating Machinery

• Automotive Power Train Modeling

• Multiphase Flow Modeling

• Comprehensive CFD Capabilities

• Special Material Processing

• Summary

Fluid Dynamics Themes


Rapid & Robust Meshing

Top on the list of challenges engineering companies are facing is shortened product development schedules while at the same time the product designs themselves are becoming increasingly complex.

Meshing of these designs introduce challenges in terms of speed, robustness and accuracy.

Courtesy Siemens AG.


Rapid & Robust Meshing

Enhanced productivity through increased automation, flexibility, efficiency and robustness

• Assembly Meshing (Tet and CutCell)• Performance (Speed, Robustness)• Selective Meshing• Virtual Topologies• Hex Meshing• ICEM CFD/TGrid


Workflow & Usability

Simulation departments are looking for improved usability and the ability to glean more information from fluid dynamics simulations for all users –from occasional designers to experienced analysts – from geometry creation through post-processing.


Workflow & Usability – Geometry Advances

Focus on enhancing your productivity through new features, increased flexibility, efficiency and usability• ANSYS DesignModeler– Core modeling improvements– Application‐specific modeling

• ANSYS SpaceClaim Direct Modeler– Improved Workbench integration– Enhanced Model Preparation

• Interoperability– Support for new CAD releases – New CAD file readers


• Workbench Design Exploration and Optimization for increased understanding, innovation & simulation ROI

• New at 14.0• Reduced time required

• 2 new adaptive DOEs • Distributed solve• Design point sorting

• Increased Robustness• Reserved licensing

• Support for partial DOE’s

• Increased understanding• New charts• Improved GOF• Project report

Workflow & Usability – ANSYS DX

Single Physics

Multiphysics

“What if” Study

Design Exploration

Optimization

Robust Design

R14


ANSYS Remote Solver Manager at R14 (FL and CFX)

• Major improvements for fluids as well as general usability and robustness improvements to RSM

• Queue multiple jobs on a local machine– Overnight or other low‐usage times

• Submit jobs to remote machines– Distributed clusters– Management of files, including UDFs

• Update design points in parallel via RSM

Workflow & Usability – ANSYS RSM


Workflow & Usability – ANSYS CFD

Improved workflow and usabilityin ANSYS Workbench • Extended User Preferences (FL)– General options– Launcher options

• Extended parameters for FLUENT– Real and profiles variables in zone and domain settings

– Examples:• Phases: Nucleation Rate, Coalescence and Breakage Kernels• Phase Interaction: Surface Tension Coefficients, Lift Coefficient, Restitution Coefficient

• Contact Angles (on wall BC)



Process more information per simulation (FL)• Monitor more than a single lift, drag, or moment monitor– Multiple monitors for each type

• Assign any variable to a custom field function for use with unsteady statistics

Export reduced data for faster file write and reduced storage needs• FLUENT – on a per cell zone basis to CFD Post, EnSight, and FieldView

• CFX ‐ Output selected solution data only at boundaries during transient β

Multiple force monitors Custom field functions for statistics

Output control panel in CFX



Enhanced ability to use ANSYS CFD with other tools• CGNS library version upgrade to 3.1• Support face‐based regions in CFD‐Post (in addition to node‐based)– Enable adoption use as common CFD post‐processor for 3rd party tools

Control of default launcher working directory (CFX) β

• Ability to switch default to directory in which launcher was started

Improved licensing preferences (CFX)• Adherence to global license settings

Example CGNS results from 3rd

party tool with face-based region definition



Easier specification of animation resolution for high definition playback (CFD‐Post)• Extended output options to include HD standard sizes– HD Video 720p– HD Video 1080p

• Also: Automatic internal adjustment to avoid Windows Media Player playback limitations


Multiphysics and Systems Coupling

Engineers need to accurately predict how complex products behave in real-world environments with real world physics.

ANSYS is continuously improving the ease of use and efficiency of simulating real world interactions between fluid dynamics, structural mechanics, heat transfer, and electromagnetics within a single, unified engineering simulation environment using systems coupling.


Two‐way surface force/displacement coupling between FLUENT and Mechanical via Systems Coupling• Steady/static and transient two‐way FSI• Integrated post‐processing with CFD‐Post• Workbench based setup and execution

– Windows and Linux• Alternative execution from command line

– including cross‐platform• Parallel processing with ANSYS HPC

– RSM currently not supported• Restarts for fluid‐structure interaction• Parameterization, design exploration and optimization

Multiphysics and Systems Coupling – FSI

Non‐Newtonian blood flow through a three leaf mitral valve


Multiphysics – 1‐way FSI

Significantly faster surface mapping for 1‐way FSI (CFD‐Post) β

• New Octree mapping method significantly faster algorithm – Need to set Option in CFD‐Post

• 1‐way FSI in ANSYS Workbench uses CFD‐Post ‘under‐the‐hood’– Will use mapping option set by user in CFD‐Post (which is stored in user preferences)

– Status message with diagnostics report indicates new mapping method is being used


Multiphysics – FLUENT‐Maxwell‐Mechanical

Complex multiphysics modeling• New: Electromagentic‐thermal interactions inside Workbench using FLUENT with Maxwell– One‐way and two‐way β coupling

• Combine with 1‐way FSI

β


Solver and HPC Performance

Companies need to make informed development decisions about their increasingly complex products in increasingly shorter time frames.

Cou

rtesy

Sie

men

s


Solver and HPC Performance – Adjoint Solver

Adjoint Solver for FLUENT fully tested, documented, and supported at R14.0

• Provides information about a fluid system that is very difficult and expensive to gather otherwise

• Computes the derivative of an engineering quantity with respect to inputs for the system

• Engineering quantities available– Down‐force, Drag, Pressure drop

• Robust for large meshes– Tested up to ~15M cell

Shape sensitivity to down‐force on a F1 car

Shape sensitivity to lift on NACA 0012

Lift Force (N)Geometry Predicted ResultOriginal ‐‐‐ 555.26Mod. 1 577.7 578.3Mod. 2 600.7 599.7Mod. 3 622 621.8


Solver and HPC PerformanceMesh Morpher and Optimizer

FLUENT Mesh Morpher and Optimizer• Create your own optimization functions using parameters

• Faster mesh deformation• Constrain some boundaries and allow others to deform

• More easily assess the effectiveness of the optimization routine– Save and plot the value of the objective function as a function of design iteration number

• Execute TUI commands before and after the optimization loop

Objective function definition panel

Example data showing faster mesh deformation

Objective function as a function of design stage



Focus on robustness, accuracy, and efficiency

• Improved performance (FL)– Out of the box performance

• New defaults– 2nd order for some equations– hybrid initialization

• Solver optimizations– Higher order term relaxation (HOTR)

• More aggressive initial settings• Improved convergence

– Migration manual

• Beta: NRBCs with PBNS (FL)– Compatible with combustion models

RAE‐2822 AirfoilM=0.73AOA=2.8 deg

Standard @ CFL=200

HOTR @ CFL=200

Standard @ CFL=100

HOTR @ CFL=100

Normalized scales residuals comparing HOTR to standard relaxation at different CFL settings

Diverges



Faster convergence on stretched meshes

• Convergence Acceleration for Stretched Meshes (CASM) with the density‐based implicit solver

• Recommended for:– Steady‐state simulations– Anisotropic meshes with high stretching in the local flow direction

– Stretched meshes with Y+ near 1

CdStandardCASM

Convergence acceleration for stretched meshes requires nearly 10x fewer iterations in this case

(300 vs. 3000 iterations)

RAE-Wing-body 5M=0.8AOA=2.0 deg



Improved scalability (FL)• Scalability to higher core counts• Simulations with monitors including plotting and printing

• Cluster‐to‐cluster view factor file writing optimization

Hex‐core mesh, F1 car, 130 million cellsmonitor‐enabled

0

5

10

15

20

25

30

35

0 200 400 600 800 1000

Example data for scaling with R14 monitors

3072 cores

Sample cluster‐to‐cluster view factor data writing using 32‐way parallel and Infiniband.

0.4 million surface clusters

1.1 million surface clusters

R13R14



Faster auto‐partitioning (FL)• Optimized for multi‐core clusters • All simulations benefit • New default• More constant time required with fixed overhead

Improved usability (FL)• Better error tracking• Latest Platform and Intel MPI versions on all platforms

Work in Progress: GPU investigation• R14.0: Viewfactor and ray tracing calculations on GPUs (FL)

Example viewfactor calculation times for different combinations of GPUs and CPUs



Optimize parallel partitioning in multi‐core clusters (CFX)β

• Partitioner determines number of connections between partitions and optimizes part.‐host assignments

Re‐use previous results to initialize calculations on large problem (CFX) β

• Large case interpolation for cases with >~100M nodes

Clean up of coupled partitioning option for multi‐domain cases (CFX)• Eliminates ‘isolated’ partition spots

Compute Node 1 Compute Node 2

P1

P5

P3

P6

P2 P7

P4 P8

P1P5

P3

P6P2

P7

P4

P8

Partitioning step finds adjacency amongst partitions; partitions with max adjacency are grouped on same compute nodes


Rotating Machinery

Rotating machinery plays a key role in many industries, including aerospace, power generation, automotive, marine, HVAC and healthcare. Manufacturers are currently challenged to improve the performance of their machines, more than ever before.

Courtesy Siemens AG


Rotating Machinery

Highly efficient time accurate simulations with Transient Blade Row capability (CFX)• Several models available– Time Transformation (TT)

• Inlet Disturbance • Single Stage TRS

– Fourier Transformation (FT)• Inlet Disturbance • Single Stage TRS β

• Blade Flutter β

Surface pressure distribution (top) and monitor point pressure (left) from an axial fan stage:

Equivalent solution with Time Transformation at fraction of computational effort

Reference solution without a TBR method, requiring 180 deg model

Time Transformation solution, requiring only

3 stator and 2 fan blades


Rotating Machinery

Workflow extensions for Transient Blade Row (CFX)• Incorporation of TBR methods in Turbo setup mode– Switch to general mode for case‐specific details

• Profile replication and clocking for inlet disturbances– Automatically ensure complete circumferential overlap of profile with mesh

– Select profile and specify rotation rate


Rotating Machinery

Integrated Transient Blade Row analysis (CFD‐Post)• Full integration of single passage TBR analysis– Each plots can be at selected time or phase position in period being modeled

– Expansion of results (to show multiple passage simultaneously) possible with macro provided with installation β


Rotating Machinery

Additional integrated turbo analysis capabilities (CFD‐Post)• Reduce need to export data and manipulate externally

• Directly assess stream‐wise changes in span‐wise distributions of circumferential averages– Look at differences or ratios of existing variables, e.g. pressure ratio

• Further derived data possible with ability to define separate lines

Easier set‐up of monitor points for rotating machinery (CFX) β

• Specification in cylindrical coordinates


Rotating Machinery

Highly automated blade row meshing – without sacrificing quality (TurboGrid)• Automatic Topology and Meshing (ATM) method expanded to handle (single) splitter blades

• Additional enhanced templates available


Automotive Power Train

With government pressure and market demand for the better fuel efficiency and reduced environmental impact of vehicles, there is a strong need to improve simulation technology for applications such as hybrid/electric powertrains, turbo chargers, direct injection, etc..


Automotive Power Train – IC Engine System

New for R14: IC Engine Analysis System provides unified and simplified set‐up for internal combustion engines• Controls simulations based on user inputs– Geometry decomposition– Meshing– Dynamic mesh set‐up– Reporting

• Parametric updates• Pervasive use of crank angle• Option to include journal files– Automate combustion and spray set‐up


Automotive Power Train – IC Engine System


Automotive Power Train – After‐Treatment

Automotive after‐treatment (FL)• Selective Catalytic Reduction (SCR)– Urea droplets and particles material– Improved DPM droplet vaporization and boiling

– Uniformity index for post‐processing

• Catalytic Converter Light‐Off– Non‐equilibrium thermal model for porous media predicts fluid and solid temperatures separately

– “Dual cell approach” automatically creates connected mesh for solid zone

– Include surface chemistry

SCR Simulation: Aqueous urea droplets and ammonia mass fraction at catalyst inlet

Plane showing gas temperature

Plane showing porous media temperature

Light‐Off Simulation: Fluid and solid temperatures for non‐thermal equilibrium porous model


Automotive Power Train – Solver meshing

FLUENT MDM improvements driven by internal combustion engine needs • Retain and remesh boundary layers during tetrahedral remeshing (FL)– Boundary layer settings from original mesh

– Example applications: internal combustion engines and FSI

• Improved robustness and usability for dynamic mesh (FL) – Mesh smoothing– Cut cell remeshing– Parallel

T=0 T=25 T=50

Tetrahedral mesh with boundary layers after remeshing

Remeshing a tetrahedral mesh with boundary layers during a simulation


Automotive Power Train – Battery Model

FLUENT Battery Model: a pragmatic approach to full‐scale battery computations• Temperature effect on battery dynamics

• Appropriate for various types of batteries

• Robust numerical algorithm for fast convergence

• User‐customizable properties and correlations

Discharge curves for a prismatic Li‐ion cell

Temperature Current DensityContours on a prismatic Li‐ion cell


Multiphase Flow Modeling

Many industrial processes involve the simultaneous flow of multiple phases.

Most of these processes are impossible to observe directly. Therefore, engineers rely on models and experiments to gain insight into improving the efficiency, throughput, safety and reliability of their processes.

Courtesy of Petrobras


0.00E+00

5.00E‐04

1.00E‐03

1.50E‐03

2.00E‐03

2.50E‐03

0 1 2 3 4 5

Dp (m

)

time (s)

Diffusion‐controlledConvection Diffusion‐controlledExperimental Data – symbols

Multiphase – DPM

Accuracy and robustness improvements

• Accurate heat and mass transfer for low and high evaporation rates in the same simulation (FL)– Convection/Diffusion controlled vaporization

• More accurate modeling for high Weber number sprays (FL)– Stochastic Secondary Droplet Breakup Model (SSD)

• Improved tracking for DPM with moving and deforming meshes (FL)– Particles are tracked in the cells deformed/moved each time step

Single droplet experiment of Wong‐Lin

Hiroyasu spray bomb, p = 5 MPa


Model dense particulate flows with DEM (FL)• DEM enabled as a collision model in the DPM model panel

• Use in combination with single phase and DDPM simulations

• Works in parallel• Particle size distributions• Prediction of the packing limit• Head‐on collisions• Collisions with walls• Example applications: Bubbling and circulating fluidized beds, particle deposition in filtering devices, particle discharge devices (silos)

Multiphase – DEM

3% fines start‐up ‐15 sec

and 15‐30 sec

Note that channeling is observed in the 15‐30 sec animation

NETL Fluidized Bed Simulations using DEM with DDPM

12% fines0‐25 sec


Multiphase – DPM

Easily replicate particle or droplet injection at different locations or in different directions (CFX)• Specify local coordinate frame for each Particle Injection Regions (PIR)

• Much more convenient for set‐up of large numbers of PIRs– Applications with multi‐port fuel injection, spray dryers, scrubbers, etc.

Directly specify swirling injection at PIRs (CFX)• Use CEL to flexibly define cylindrical components as f(position, time, …)

• Includes extension to LISA model for pressure‐swirl atomizers


Multiphase – DPM

Generate additional particle data output for further analysis (CFX) β

• Frequency distributions of particle quantities at boundaries to *.csv for histograms in CFD‐Post or elsewhere

• Wall particle monitoring per patch– Instantaneous and time‐integrated

Extend models with algebraic AVs for particles (CFX) β

• Analogous to fluid AVs

Enhanced FLUENT particle track post‐processing (CFD‐Post)• Vector variables and sizing by diameter

Time-integrated mass of particles

hitting a wall

Histogram of particle size distribution at an outlet


Multiphase – Condensation

Ability to include global effect of wall condensation without multi‐phase details (CFX)• Single phase, multiple components– Mixture of one condensable and one or more non‐condensable species

• Condensable component extracted by sink terms at walls and CHT boundaries, as function of concentration through boundary layer– Liquid film is not modeled

• Key application: nuclear accident scenarios looking at containment pressure variation over time need to include macroscopic effect of condensation

Water condensation near a wall at a fluid-

solid interface


Multiphase – Free Surfaces

Better robustness and faster convergence for free surface steady‐state cases using coupled VOF (FL)• Improved in R14.0

Options for better trade‐off between stability and accuracy (FL)• Hybrid treatment of Rhie‐Chow face flux interpolation with special treatment near free surfaces

Coupled VOF after 500 iterations

Coupled P‐V, Segregated VOF after 1400 iterations


Multiphase – Free Surfaces

Continuum surface stress method for surface tension flexibility (FL)• Surface tension as a function of any variable

• Alternative to existing CSF model in some difficult cases– For example: cases with sharp interface discontinuities Sample data: surface tension and viscosity

as a function of temperature

Experiment image compared to FLUENT results. Image from: Thermally Induced MarangoniInstability of Liquid Microjets with Application to Continuous Inkjet Printing by Furlani et al.


Multiphase – Eulerian

Boiling model extensions, and testing (FL)• Critical heat flux (CHF) for modeling boiling dry out conditions

• Transition smoothly between the bubbly and droplet regimes

• Boiling with bubble size distributions using interfacial area concentration (IAC) models• Accurate interfacial areas for heat and mass transfer calculations in non‐equilibrium boiling conditions

• Example applications: Nuclear industry, engine jacket cooling

Axial Position (m)

Wall Tem

perature

Test case based on the data in Hoyers et. al. showing dry out at the wall

IAC model allows bubble diameter variation at different stations along the pipe to be modeled. Data from: Bartolemei et. Al



Combine phase change and varying dispersed phase size distributions for improved accuracy (CFX) β

• Homogeneous and inhomogeneous MuSiG can be applied together with general phase change and RPI wall boiling models

Simulation and comparison with experimental results from DEBORA Test Facility, courtesy of HZDR (formerly FZD)

= +


Multiphase – Population Balance

Model diameter and other variable changes (bin fraction, moments) due to density changes in the dispersed phase (FL)• Compressible dispersed phase• Example applications: Geophysical flows, oil and gas flows, compressible flows with population balance

Model growth and nucleation with the inhomogeneous discrete model (FL)• Example applications: Crystallization, bubble columns with mass transfer

Effect of expansion on bubble diameter in bubble column with discrete method: Monodispersedbubbles injected at bottom and results compared with analytical diameter (in white)


Multiphase – Population Balance

DQMOM Population Balance captures the segregation of poly‐dispersed phases due to differential coupling with the continuous phase (FL) • Faster solution time than the inhomogeneous discrete model

• Multi‐fluid model convects different dispersed phase sizes using different velocities

• Example applications: Fluidized beds, gas solid flows, spray modeling, bubble columns

Velocity big bubbles > velocity small bubbles

All bubbles move with same velocity

DQMOM QMOMg

Bubble diameters for DQMOM (white) compared to QMOM (red)

Contours of volume fraction of the phase with the largest diameters in a bubble column.


Multiphase – Eulerian Wall Film

Eulerian Wall Film Model for rain water management, deicing and other applications • Available models:– Momentum coupling– DPM coupling

• Particle collection, splashing and shearing

– Heat transfer

Contours of temperature in an Eulerian Wall Film with heat transfer case

film

Q = 3000 W

airsolid

liquid water injected

KCmQT

Pfilm 75

400001.03000

The wall film on a car mirror withdroplets released due to wind shear


Comprehensive CFD Capabilities

Manufacturing companies today face many challenges, ranging from increased product complexity to tightened quality requirements to yield and productivity pressures.

As product complexity increases and the margin for error decreases, CFD must rise to meet growing demands for comprehensive and advanced product capabilities.

Cou

rtesy

Bor

g W

arne

r Tur

bo &

Em

issi

ons

Sys

tem

s.

Cou

rtesy

MTU

Aer

o E

ngin

es G

mbH

.


Comprehensive CFD – Motion

Improved accuracy with simplicity of immersed solids (CFX)• Addition of boundary model for more realistic velocity forcing with immersed solids– Track nodes nearest to immersed solid– Assume constant shear (laminar) or use scalable wall function (turbulent) to modify forcing at immersed solid ‘wall’

• Can improve immersed solid predictions significantly– Continuing development for further improvements and broader applications

Boundary-fitted mesh

Immersed Solid (default)

Immersed Solid with Boundary Model


Comprehensive CFD – Motion

Further user control on deforming mesh to allow retention of optimal mesh quality (CFX) β

• Diffusion of boundary mesh motion can be defined to be anisotropic – i.e. preferential diffusion in different directions

Sample case: Deformation of an originally square domain xy xy

Final mesh with isotropic diffusion: skewed elements

Final mesh with anisotropic diffusion: improved mesh quality, greater range of motion possible


Comprehensive CFD – Interfaces

Easy simulation of opening and closing (CFX)• Conditional GGIs that can open and close as the solution progresses

• Define condition as CEL function of solution– e.g. After set time, based on solution values (single or integral values)

• Example applications– Membranes bursting, windows shattering, valves opening/closing

• Define as reversible (e.g. valve) or irreversible (e.g. burst membrane)


Comprehensive CFD – Dynamic Mesh

Greater accuracy and speed with default interface settings (CFX)• Direct intersection method is now default – Significantly faster

• biggest impact on transient cases with re‐intersections

– More accurate than bitmap method

Support for dynamic mesh on meshes with polyhedral elements (FL)• Smoothing on polyhedral elements • Polyhedral elements skipped during remeshing


Comprehensive CFD – Turbulence

Focus on wall treatment accuracy (FL)• More accurate rough wall treatment for epsilon‐based models – Avoids reduction in roughness when the near‐wall mesh is refined

– New default for all simulations using rough wall treatment

• Reduced sensitivity to the near‐wall mesh density for the Spalart‐Allmaras model– New enhanced wall treatment is the default for S‐A turbulence model

Turbulent flow over a flat plate with roughness and heat transfer

R14 wall treatment R13 default treatment

Sensitivity of the skin friction coefficient to mesh density in an incompressible flat boundary layer modeled with Spalart‐Allmaras



Improved accuracy for turbulent flows with strong rotation and streamline curvature with one and two‐equation models (FL)• Option to apply a correction term sensitive to rotation and curvature

• Accuracy comparable with the RSM, with less computational effort

Improved accuracy for turbulence near porous jump interfaces (FL) β

• Use wall functions to include the effects of solid porous material on the near‐wall turbulent flow on the fluid side of porous jump interfaces

Contours of curvature correction function (fr) in a curved duct (k‐omega SST turbulence model)

Contours of velocity showing the impact of a porous jump on velocity in bordering cells



More accurate solution of high Re wall bounded flows using LES (FL)• Algebraic Wall Modeled LES (WMLES) formulation based on Smagorinsky model

• Benefits gas turbine combustors and other internal flow applications

Apply scale‐resolving turbulence models only locally, as required, to balance cost vs. accuracy (CFX) β

• Zonal LES model using turbulence forcing

• Forcing zone defined by logical CEL function

Flow over a wall mounted hump

Contours of Q criterion

A mixing layer with resolved turbulence using SAS initiated by the forcing model


Comprehensive CFD – Acoustics

Validations and model extensions (FL)• Convective effect for FW‐H acoustics solver– Option to include the effect of far‐field velocity on the generated sound for the Ffowcs‐Williams & Hawkins solver

– Improves accuracy when modeling aeroacoustics and external flows

• Model Doppler effects due to the relative motion of acoustic sources and receivers (FL)– For example: Sound from a source moving with a constant speed (airplane, car)

M=0.2

FWH source surface,R=1m

Receivers, R=3m

2D monopole with convection

OASPL – overall sound pressure level

M=0.4Moving receiver

SourceSound pressure compared with analytic solution at approach and departure (Doppler effect different frequencies)

Acoustics directivity


2x2 3x3 6x6

Serial/Old 218.63 s 269.95s 536.32s

Serial/New 202.40s 241.56s 429.40s

Parallel 2‐proc/Old

86.4s 107.4s 205.3s

Parallel 2‐proc/New

80.7s 98.4s 167.2s

Comprehensive CFD – Radiation

More Efficient Discrete Ordinates Radiation (FL)• DO radiation calculations ignore solid zones not participating in radiation or participating in heat transfer by conduction only

• Avoids unnecessary CPU time and memory allocations• Performance improvement varies depending on specifics of case– Up to 20% improvement for test cases

Sample data for a 3D case with 5 solid zones


Comprehensive CFD – Shell Conduction

Shell conduction: Improved accuracy and ability to include combustion (FL)• Non‐premixed and partially‐premixed combustion models

• Example applications: Gas turbines, those with thin walls and combustion

Improved shell conduction usability (FL)• Updated documentation for wall temperature variables and shell zone and thin wall post‐processing

• Post‐process the external wall temperature if shell conduction is applied on a one‐sided‐wall

Oxidizer(air)V=0.1 m/s, T = 300KMean Mixture Fraction=0

Fuel, V=1 m/sT= 300KMean Mixture Fraction=1

External Wall with convective BCh=2, Tinf= 300K

Temperature on a mid‐geometry surface

Geometry and details for shell conduction with non‐premixed combustion test case


1D Reacting Channel Model (FL)• Detailed chemistry in the channel (plug flow)– Simple geometry, complex chemistry

• 3D heat transfer in the furnace/shell – Complex geometry,simple (or no) chemistry

• Example applications: cracking furnace, fuel reformers, …

Comprehensive CFD – Reacting Flows

1D reacting channel model in FLUENT

1D reacting channel geometry in FLUENT



Focus on Validation and Verification (FL)• Real gas – RCM1, RCM3

• Surface chemistry – Kleijn CVD

• Turbulent non‐premixed flames – Sydney bluff body flame

• Turbulent premixed flames– Chen F3

• Internal Combustion Engines

Example simulation results for the Kleijn CVD verification for surface chemistry


Comprehensive CFD – Combustion

More accurate, flexible, and efficient combustion simulations with CFX‐RIF (CFX) β

• Variable thermodynamic conditions• User‐defined mechanisms• Reduced materials list

Expanded modeling with inert residual material model (CFX) β

• Application example: Safety analyses involving hydrogen combustion together with fire suppression using water


Comprehensive CFD – Combustion

Faster, more accurate, and more general unsteady combustion (CFX) β

• G‐equation model upgrade– GUI exposure– Robustness enhancement– Additional correlations for flame speed and flame thickness

– Improved compatibility with ignition models

– More accurate and faster flame front re‐initialization

Easy customization of non‐ideal mixture properties (CFX) β

Flame front prediction from G-equation shortly after ignition in an IC engine


The polymer, glass, metals and cement processing industries face the special challenge of highly complex material behaviour when performing simulations to minimize physical prototyping in the manufacture of extrusion dies or improve the quality of thermoformed or blown products.

Importantly, for say the packaging industry, is the ability to predict more accurately the structural performance of the package in real-use scenarios.

ANSYS Polyflow provides a host of special material models and enhanced capabilities to meet these needs.

Special Material Processing


• Polydata will support named boundaries and subdomains, using Named Selections in ANSYS Workbench

• ANSYS POLYFLOW supports 1‐D and 2‐D pmeshes generated in ANSYS Meshing and ANSYS ICEM CFD– 1‐D pmeshes for 2‐D meshes

– 1‐D and 2‐D pmeshes for 3D

Enhanced Workflow and Usability in Workbench


• A database is now available of ANSYS Workbench Project Templates for common applications of ANSYS Polyflow– Templates folder in the POLYFLOW 14.0 installation

• Each project contains a complete calculation (Geometry to Post‐processing) with parameters and a final report– Blow‐molding, extrusion, thermoforming examples in ANSYS Workbench archived (.wbpz) format

• The template projects can be used for learning purposes or as starting points for individual projects.

Enhanced Usability in Workbench from Templates


• ANSYS POLYFLOW 14.0 can import and run Cutcellmeshes for complex extrusion dies– New mesh generation method in Tgrid and ANSYS Workbench

– Fast and efficient algorithm for rapid hex dominant mesh generation on highly complex geometry, such as extrusion dies

Enhanced Extrusion Solutions from use of CutCell Meshes

• A “Sliceable” mesh can be created for the free jet region from a CutCell mesh, to allow computation of final extrudate shape

• Creates conformal prismatic mesh in the free jet subdomain with CutCell mesh in the die

• Take full advantage of rapid hex dominant CutCellmeshing for complex die geometries

Initial CutCell mesh

Sliceable free jet mesh

Final extrudate shape


• ANSYS POLYFLOW will now allow numerical decoupling of species and VOF volume/fill fractions from the flow calculation– Reduce memory requirements and improve solution speed by segregating the volume fraction field

• This is very useful for applications like multi‐fluid co‐extrusion (and glass forming)

Enhanced Extrusion Solutions from Numerical Decoupling of Species and VOF

Fill fraction of the primary fluid in multi-fluid co-extrusion: Flow through a complex die modeled with VOF method


• A force can be specified for mold motion with shell meshes for parisons in blow molding and thermoforming

• Specify maximum displacement of the mold to set a bounding criterion for the motion

Enhanced Solutions for Blow Molding and Thermoforming

• A new contact release method has been implemented for blow molding and thermoforming

• This allows determination of Parison “peeling” from mold for blow molding and plug release in thermoforming

• Contact release is compatible with both specified force and specified velocity for the mold

Thermoforming of a medical blister


• Thickness data from ANSYS POLYFLOW can be sent to ANSYS Mechanical and ANSYS Explicit Dynamics via an automatic connection

• Complete Virtual Prototyping and Testing capability in ANSYS Workbench for packaging manufacturing:– Simulate blow molding or thermoforming process to get final thickness distribution with Polyflow

– Based on Polyflow predicted variable thickness, perform stress and deformation analysis• Top load stress analysis with ANSYS Mechanical shown here

Virtual Prototyping for Packaging Applications ‐Mechanical


• Where the deformation or loading rate are larger, then transfer the Polyflow thickness data to Explicit Dynamics system in Workbench– Explicit Dynamics is available as part of ANSYS Explicit STR and ANSYS AUTODYN

• Here we will load a blown part (created from Polyflow) with a rigid plate (in green in mesh plot)– Loading normal to the plate

Virtual Prototyping for Packaging Applications – Explicit Dynamics


ANSYS Fluid Dynamics 14 contains enhancements in all products…• ANSYS FLUENT, ANSYS CFX, ANSYS CFD‐Post• ANSYS Turbogrid• ANSYS Polyflow

… for rapid and robust meshing, workflow and usability, HPC and solver speed, rotating machinery, automotive and multiphase applications as well as overall to provide comprehensive CFD.

Summary


Questions?


Appendix



New mesh options in FLUENT

• Copy and move zones (TUI)– Copy zones and translate/rotate the copies directly in FLUENT

• Preserve interior faces during conversion to polyhedral meshes (TUI) – Keep interior faces that may be needed for post‐processing

Copied and translated zones in FLUENT 14


FLUENT 14 Solver Numerics

Embedded Reference Frames: Local frame specification for sliding mesh• Local frame (origin and axis) defined with respect to moving (base) frame

• Origin and axis locations remain fixed with respect to base frame

• NO UDF required to update the origin or axis!

Local (embedded) frame

Base frame

Local frame set to revolve about an origin at center (0.1, 0.1) w.r.t. base frame



Accelerated convergence for complex calculations• Equation specific control for pseudo‐transient calculations (FL)– Accessed in the expert panel

PBCS – Pseudo Transient off PBCS – Pseudo Transient fully on PBCS – Pseudo Transient on for all equations except species and energy

244 iterations

125iterations

66iterations



Work in Progress: GPU investigation• Acceleration limited to Jacobi/Gauss‐Seidel smoothers in the context of flexible cycle on fine level

• Special interleaved CSR matrix storage structures for faster memory access on GPUs

• Coloring algorithm for parallelizing Gauss‐Seidel on GPU

R14: Viewfactor calculations on GPUs as beta (FLUENT)

Research continues…

X X X X X

X X X X X

X X X X X X

X X X X X

X X X X X X

X X X X X

X X X X X X

X X X X X X

X X X X X X X

X X X X X



Improved numerical stability for user‐defined mass transfer with compressible phases (FL) • Linearized option for user defined mass transfer also allows larger timesteps

New Hulin‐Gidaspow and Gibilaro drag laws for granular flows (FL)• Modified Huilin‐Gidaspow drag allows smooth transition from dilute to dense limit, improving numerical behavior

• Gibilaro drag law for fluidized beds provides a better representation of physics in these cases

Cavitation in a nozzle using linearized mass transfer

vof of liquid

Absolute pressure



More Turbulence Enhancements (FL)

Time statistics data sampling time displayed in GUI

Turbulence dissipation rate () variable for omega‐based turbulence models available in CFD‐Post

Reynolds stress normal components accessible via UDF• Eliminates need to compute Reynolds stress normal components from scratch in a UDF



Minor improvements (FL)• Surface chemistry robustness• UDF hooks for PDF table• Tabulated laminar flame speeds in partially‐premixed PDF table for non‐adiabatic

• Chemically‐Activated bi‐molecular reactions

• Local diffusion for spark model• Real gas with partially‐premixed model


FLUENT 14 DPM (beta)

Node based averaging of DPM source terms and DDPM volume fraction • Standard averaging dumps all volume fraction into one cell.

• Node based averaging distributes volume fraction over several cells by collecting data on mesh nodes.

• Several kernels available.• Strongly reduces grid dependency.• Improves convergence behavior for steady simulations.

• Allows for larger time steps in transient simulations.

• Requires more memory.

0.013

0.0

volume fractionstandard average

volume fractionnode based average


Rotating Machinery

Time Transformation

• Solve in transformed ‘inclined’ time, fully implicit, with turbulence

• Results in inclined time stored using Fourier‐based data compression

• Results re‐constructed in physical time for analysis

Fourier Transformation

• Fourier–series used for reconstruction of solution history on pitch‐wise boundary and inter‐row interfaces for efficient data storage & convergence

• Two passages for better signal collection and faster convergence than single passage

ansys 14.0 fluid dynamics update - dipankar choudhury

Documents

courtesy siemens ag

xfinal mesh

workflow

rapid

immersed solid

base frame

improve